基于松弛策略的文本层次分类方法.PDF

第43卷 第8期 北 京 工 业 大 学 学 报 Vol.43 No.8 2017年 8月 JOURNAL OF BEIJING UNIVERSITY OFTECHNOLOGY Aug. 2017 基于松弛策略的文本层次分类方法 杜永萍,杜晓燕,陈守钦 (北京工业大学计算机学院,北京摇 100124) 摘摇 要:为了进一步提高文本层次分类的性能,在传统层次分类方法的基础上融入了松弛策略思想,在构造层次结 构的过程中,该方法推迟了不确定类别的节点判定,直到可以明确所属类别,大大降低了高层节点分类错误对低层 节点分类性能的影响,即“阻滞冶问题的有效缓解. 实验结果表明:松弛策略思想可以构建更加合理的层次结构,并 进一步提高了分类的性能;相对于支持向量机等其他分类方法,在时间性能上更加高效,对于大规模文本分类任务 而言具有重要意义. 关键词:松弛策略;层次分类;支持向量机;聚类 中图分类号:TP391 文献标志码:A 文章编号:0254-0037(2017)08-1175-07 doi:10.11936/ bjutxb2016040059 Relaxed Hierarchy Structure Construction for Text Classification DU Yongping,DU Xiaoyan,CHEN Shouqin (College of Computer Science,Beijing University of Technology,Beijing 100124,China) Abstract:Hierarchical classification is an effective method to solving the classification problem on the large鄄scale text data and it can save timewithout reducing the classification accuracy. A relaxed strategy combining the traditional hierarchical classification method to improving the system performance was introduced. During the process of hierarchy structure construction,the nodejuedgement of the uncertain category was delayed until it was classified clearly. The “block冶 problem was effectively alleviated to transfer the classification error from the higher level to the lower level in the hierarchy structure. The experimental results show that the relaxation approach can build a more reasonable hierarchy and further improve the classification performance. Compared to the other classification method, such as Support Vector Machines,themethodhasmoreadvantageintimeperformanceandismoreefficientforlarge鄄scale text classification task. Key words:relaxed strategy;hierarch